Prospective bidirectional associations between depression and chronic kidney diseases

Study design

The CHARLS study used a multistage clustering sample method to select participants from 28 provinces in mainland China²³. A total of 17,708 participants from 10,257 households recruited from 28 provinces within China were included at baseline (2011–2012, Wave 1). CHARLS respondents are then followed up every 2 years, using a face-to-face computer-assisted personal interview. Two subsequent follow-ups were carried out in 2013–2014 (Wave 2) and in 2015–2016 (Wave 3), respectively. At each wave, trained personnel conducted face-to-face interviews to collect information including participants’ sociodemographic characteristics, medical history, health behaviors, and measures of general health status and functioning including depressive symptoms. More detailed description of CHARLS can be found in previous publications^23,24. This study was approved by Biomedical Ethics Review Committee of Peking University, and all participants signed informed consents. The ethics application for collecting data on human subjects in CHARLS was approved by the Biomedical Ethics Review Committee of Peking University (IRB00001052-11,015). All study methods were carried out based on the Declaration of Helsinki.

Study sample

In study I, individuals who met all of the following criteria were included: aged at least 45 years, reported information about the Center for Epidemiological Studies Depression Scale-10 (CESD-10) in Wave 1, without CKD diseases in Wave 1 and reported information about CKD in Wave 3. Finally, a total of 7637 individuals were eligible for subsequent analysis (Fig. 1HAS). In study II, individuals who met all of the following criteria were included: aged at least 45 years, reported information about CKD and in Wave 1, without depression in Wave 1 and reported information about the CESD-10 in Wave 3. Finally, a total of 6337 individuals were eligible for subsequent analysis (Fig. 1B).

Flow chart of the selection of the study population for Study I (HAS) and Study II (B).

Depression, CKD and covariate assessments

The CESD-10 was administered to measure depressive symptoms at baseline and each follow-up visit of the CHARLS²⁵which has been proven to be a reliable and valid approach to detect depression in Chinese adults^26,27. The CESD-10 scale incorporates depressed mood and positively affected parts and consists of ten items. Its sum scores range from 0 to 30, with higher scores indicating higher degrees of depressive symptoms. A cutoff of 10 was used to identify depression individuals²⁸.

CKD was based on self-reported physicians’ diagnosis ‘Have you ever been told by a doctor that you have kidney diseases?’ or personal eGFR level²⁹. If an affirmative answer was given by an individual or his/her proxy respondent or eGFR < 60 mL/min per 1.73 m², then the individual would be categorized as experiencing the first-time CKD in his/her life (except for tumor or cancer). Serum creatinine was measured using a rateblanked and compensated Jaffe creatinine method. Assessment of eGFR used the Chronic Kidney Disease Epidemiology Collaboration creatinine equation with adjusted coefficient of 1.1 for the Chinese population³⁰.

The other covariates collected included age, sex, living place (rural vs. urban), smoking status (ever smoking vs. never smoking), drinking status (ever drinking vs. never drinking), body mass index (the weight in kilograms divided by the square of the height in meters), the presence or absence of other chronic diseases (defined by a self-reported history of dyslipidemia, diabetes, cancer, heart disease, stroke, asthma, lung disease, liver disease, digestive disease, and memory trouble). “Ever smoking” means that the respondent reports smoking at some point, and “never smoking” means that the respondent reports never having smoked. “Ever drinking” means that the respondent reports having had an alcoholic beverage in the past, and “never drinking” means that the respondent reports not having any alcoholic beverage in the past. Moreover, dyslipidemia (defined by a history of dyslipidemia, or triglyceride ≥ 2.26 mmol/L, or total cholesterol ≥ 6.22 mmol/L, or high-density lipoprotein cholesterol < 1.04 mmol/L, or low-density lipoprotein cholesterol ≥ 4.14 mmol/ L), diabetes (defined by a history of diabetes, or fasting glucose ≥ 6.1 mmol/L, or non-fasting glucose ≥ 7.8 mmol/L)³¹.

Statistical analysis

Participants’ baseline characteristics are presented as percentages for categorical variables, as the means with standard deviation for normally distributed variables and as medians with interquartile range for non-normally distributed variables. Demographic and clinical characteristics were compared between groups by Student t tests, or Wilcoxon rank-sum tests for continuous variables and χ2 test for categorical variables as appropriate. Multivariate logistic regression models were used to calculate odds ratios (ORs) and 95% confidence intervals (95% CIs) of the relationships between depression and CKD. Potential covariates, such as age, sex, living place, education level, smoking, drinking, body mass index, systolic blood pressure, and medical history (dyslipidemia, diabetes, cancer, heart disease, stroke, asthma, lung disease, liver disease, digestive disease, and memory problem) were included in the multivariable models. Furthermore, subjects were further divided into three groups according to CESD-10 scores: < 10; 10–12; ≥ 12 in study I, and according to eGFR level and self-reported kidney diseases: eGFR ≥ 90 mL/min per 1.73 m²; 60 ≤ eGFR < 90 mL/min per 1.73 m²; eGFR < 60 mL/min per 1.73 m² or with reported kidney diseases in study II. The subgroup analyzes were further performed to evaluate the association between depression and CKD according to sex, age, living place, smoking, drinking, education level, history of hypertension and diabetes. In sensitivity analyses, we further adjusted baseline eGFR level in study I and baseline CESD-10 scores in study II based on multivariable models. Two-tailed P< 0.05 was considered to be statistically significant. All statistical analyzes were conducted using SAS statistical software (version 9.4, Cary, NC).